EP1094325A2 - Procédé et dispositif de détermination du nombre de sources de décharges partielles - Google Patents

Procédé et dispositif de détermination du nombre de sources de décharges partielles Download PDF

Info

Publication number
EP1094325A2
EP1094325A2 EP00660185A EP00660185A EP1094325A2 EP 1094325 A2 EP1094325 A2 EP 1094325A2 EP 00660185 A EP00660185 A EP 00660185A EP 00660185 A EP00660185 A EP 00660185A EP 1094325 A2 EP1094325 A2 EP 1094325A2
Authority
EP
European Patent Office
Prior art keywords
pulses
pulse
partial discharge
basis
characteristic parameter
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
EP00660185A
Other languages
German (de)
English (en)
Other versions
EP1094325B1 (fr
EP1094325A3 (fr
Inventor
Mats Björqvist
Vesa Latva-Pukkila
Pertti Pakonen
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
ABB Oy
Original Assignee
ABB Oy
ABB Substation Automation Oy
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by ABB Oy, ABB Substation Automation Oy filed Critical ABB Oy
Publication of EP1094325A2 publication Critical patent/EP1094325A2/fr
Publication of EP1094325A3 publication Critical patent/EP1094325A3/fr
Application granted granted Critical
Publication of EP1094325B1 publication Critical patent/EP1094325B1/fr
Anticipated expiration legal-status Critical
Expired - Lifetime legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing

Definitions

  • the invention relates to a method of separating partial discharge pulses originating from various partial discharge sources in an electric system and preferably of determining the number of the partial discharge sources.
  • the voltage strength of an insulating structure refers to its ability to endure voltage stress without electric discharges that cause disturbances or damage. If the voltage stress in an the insulating structure is increased sufficiently, discharges occur which make the insulation completely or partially conductive. The latter are called partial discharges. A partial discharge does not unite the electrodes and, thus, the insulating properties of the insulating material do not completely disappear. Partial discharges do, however, wear the insulating material and thus further weaken its voltage strength and may finally lead to a complete electric discharge. Partial discharges can be divided into two main groups, internal and external discharges. Internal discharges comprise cavity discharges and external discharges comprise surface, corona and spark discharges. Each group can further be divided into several subgroups which are often difficult to clearly distinguish from each other.
  • Partial discharge pulses are very fast pulses and usually occur as pulse groups.
  • a partial discharge and the reversal of charge that occurs in connection with it show as a current pulse in the connectors of the insulating material. In practice, these current pulses also sum into the phase voltage of the system.
  • Partial discharges can, for instance, be detected by means of an electric measurement, acoustically by the sound they generate or on the basis of the electromagnetic radiation they produce. To detect the existence of partial discharges, it is also useful to distinguish from each other the partial discharge pulses originating from different partial discharge sources. If the partial discharge pulses originating from different sources can be classified into their own groups, identifying the cause of the partial discharge becomes easier, as the partial discharge sources can be analysed one at a time and it is also possible to determine the number of the partial discharge sources. Partial discharges can be generated simultaneously in several partial discharge sources, but known solutions for detecting partial discharges usually cannot distinguish from each other partial discharges coming from different sources.
  • the invention is based on classifying partial discharge pulses originating from different partial discharge sources in an electric system into pulse groups on the basis of one or more characteristic parameters defined from them.
  • the number of the partial discharge sources can be defined by means of the pulse groups.
  • the method and system of the invention provide the advantage that they do not require any training phase performed in advance, but allow the distinguishing from each other of partial discharge pulses originating for instance from different partial discharge sources of different types or at various distances to the extent that they differ from each other within the scope of the used characteristic parameters.
  • the method and system of the invention facilitate the identification of the cause of a partial discharge by enabling the analysis of partial discharge sources one at a time and also enable the automatic determination of the number of the partial discharge sources.
  • FIG. 22 A block diagram of a method and system of the invention is shown in Figure 22.
  • the invention can be divided into two main blocks: sampling, measuring signal filtering and partial discharge pulse collection 400 and partial discharge pulse analysis 300.
  • Figure 22 also shows the operation of the analysis block 300 illustrated by means of a flow chart.
  • the use of the method and arrangement of the invention is not limited to any specific system, but they can be used in connection with various electric systems, such as electric networks or apparatuses, to distinguish from each other any partial discharge pulses originating from different partial discharge sources and also to determine the number of the partial discharge sources.
  • the arrangement of the invention can be implemented by means of digital signal processing equipment, for instance.
  • Figure 23 shows a flow chart illustrating the operation of the sampling, measuring signal filtering and partial discharge pulse collection block 400 according to an embodiment thereof.
  • a high-frequency voltage or another variable from which partial discharge pulses can be distinguished
  • the used measurement method bears no significance to the basic idea of the invention and depends, for instance, on the electric system being examined.
  • the lower limiting frequency f l is a few tens of kilohertz, for instance, and the upper limiting frequency f u is several megahertz, for instance (e.g. f l ⁇ 60kHz and f u ⁇ 8 MHz).
  • the digitising results 130 at a sampling frequency of 16 MS/s, for instance, in a total of 960,000 samples during three network cycles.
  • This three-cycle packet is in the following called a measurement and it is also the input data of an algorithm.
  • the algorithm preferably eliminates 100, 121, 200 interfering signals from the measurement using digital filtering, for instance.
  • the algorithm picks partial discharge pulses from the sample string remaining in the measurement. From the pulses, the algorithm calculates 111 for instance the following parameters depicting the pulse form: rise time of the pulse (preferably 10 to 90% points), fall time of the pulse (preferably 90 to 10% points), width of the pulse (preferably at a height of 50%). Which pulse parameters are defined at this stage depends on the characteristic parameters to be used later in the analysis stage.
  • the above-mentioned pulse parameters are stored 120 for each pulse for analysis.
  • the pulse parameters mentioned earlier are calculated only in the analysis block 300.
  • the filtering of the measuring signal and the collection of the partial discharge pulses can also be performed utilising other methods without this having any significance with respect to the basic idea of the invention. It is also possible that only a part of the presented filtering methods are used or that no filtering is needed, if the system being examined is protected against external interference.
  • Narrow-band interference refers to interfering signals whose spectrum is narrow, i.e. the energy of the signal is concentrated on a narrow frequency range.
  • Various radio transmitters and communications equipment operating on a narrow frequency range typically cause narrow-band interference.
  • This type of interference may mask weak partial discharge pulses, thus weakening the sensitivity of the partial discharge measurement.
  • Another interference type related to partial discharge measurements is asynchronous impulse interference which is a pulse-form interference and does not occur synchronously with a phase voltage (nominal frequency, e.g. 50 or 60 Hz), in other words, in consecutive phase voltage cycles, the pulses do not occur at the same phase angles. The time between consecutive interfering pulses remains almost constant, however.
  • Commutating pulses of an inverter are a typical example of asynchronous impulse interference.
  • a third interference type related to partial discharge measurements is synchronous impulse interference which is a pulse-form interference and occurs synchronously with a phase voltage. Interfering pulses repeat in consecutive cycles at nearly constant phase angles. In addition, the amplitude of the pulses remains nearly constant. Synchronous impulse interference is caused by commutating pulses of rectifiers and phase angle regulation, for instance.
  • Narrow-band interference is shown as peaks in the amplitude spectrum.
  • the width of the peak is directly proportional to the width of the interference band.
  • the power of partial discharges and noise is evenly distributed along the entire frequency range of the spectrum.
  • the amplitude spectrum of white noise is according to its specification constant on the entire frequency range. With coloured noises, the power is higher on some frequency ranges, but even these cases show as planar areas in the amplitude spectrum. In the tests performed with the measuring system used as an example in the application, the applicant has noticed that the power of partial discharges is distributed along the entire frequency range in the spectrum.
  • Partial discharges do, however, have more power at low frequencies than at high frequencies.
  • Figure 4 which discloses the amplitude spectrum
  • the peaks shown in Figure 4 are caused by narrow-band interference.
  • Narrow-band interference has in the example of Figure 4 concentrated on low frequencies, but in practice, they may occur on the entire frequency range.
  • FIG. 1 A block diagram illustrating a method of eliminating narrow-band interference is shown in Figure 1.
  • the method is based on modifying 110 a finite-length signal in a frequency domain.
  • a finite-length signal in a frequency domain.
  • the time interval of the signal After the time interval of the signal has been modified, it is returned 109 to the time domain by a reverse Fourier transformation.
  • the examples use a 60 ms-long signal time interval at a 16 MHz sampling frequency, whereby a total of 960,000 sample points are obtained.
  • the length of the signal time interval may differ from the above exemplary value.
  • Figure 2 shows the exemplary signal g(t) in a time domain, wherein the vertical axis depicts the signal amplitude g(t) and horizontal axis the time t.
  • ( Figure 3), which depicts the signal amplitude
  • Figure 4 shows the amplitude spectrum
  • the uniform area at the bottom of the amplitude spectrum can be considered the bottom level. Peaks do not belong to the bottom level.
  • the amplitude spectrum is divided into sections, for instance 32 sections, and a median is defined for each section.
  • the first of the 32 sectional medians is preferably left out, if in sampling, the measuring signal was high-pass filtered to remove the main voltage and harmonics, in which case the median of the first section does not represent the actual form of the envelope.
  • is adjusted 104 by dividing it by the envelope values sample by sample.
  • o is shown in Figure 6.
  • the bottom level is nearly constant and the peaks rising from it are narrow-band interference. Narrow-band interference peaks are easy to identify in an adjusted amplitude spectrum, if a bottom level has been defined.
  • a simple solution would be to use a fixed cut level.
  • the level When using a fixed cut level, the level should be set so high that one can be absolutely sure of not interpreting the bottom level, and, at the same time, the partial discharges, as interference. In this case, however, the sensitivity of interference elimination decreases, i.e. some of the interference is not eliminated.
  • the most advantageous solution is to define the cut level case by case.
  • the cut level must be set as low as possible, but clearly above the bottom level, however.
  • An optimal cut level can be defined by means of the average and standard deviation or variance of an adjusted amplitude spectrum, for instance.
  • the amplitude distribution of the adjusted amplitude spectrum is shown in Figure 10, wherein the horizontal axis shows the amplitude value and the vertical axis the probability density of the amplitude value. The distribution resembles an ⁇ 2 distribution.
  • ⁇ 2 distribution there is a formula for an ⁇ 2 distribution, as there is one for a normal distribution, by means of which it is possible to estimate how large a part of the values is within the given limits. For instance, of the values of a normally distributed signal, 95% is within the limits ⁇ 1,96 ⁇ ( ⁇ is the average and ⁇ is the standard deviation). The formula of the ⁇ 2 distribution is in the same form as that of the normal distribution, but the probability limits differ.
  • the cutting of the spectrum peaks is most preferably performed in two parts; first the possible high peaks are removed 106 and then the remaining low peaks are removed 108.
  • By performing the cutting in two parts produces an exact and robust result. In addition, this ensures that partial discharges are not attenuated.
  • the first cutting 106 endeavours to remove the high peaks which strongly affect the standard deviation, in which case, in the second cutting 108, the cut level level can be defined as exactly as possible above the bottom level.
  • an acceptable level value of the second cutting could be between 5 and 15.
  • the cut levels are set so high that the entire amplitude spectrum remains below them.
  • Removing the spectrum peaks from the spectrum is done at the first stage in such a manner that the frequency ranges having values exceeding the high peaks cut level 61 in the adjusted amplitude spectrum
  • the spectrum G(j ⁇ ) (like the amplitude spectrum) is discrete in relation to the frequency, i.e. it is made up of frequency samples, the smallest frequency range which can be nulled is a frequency range of the length of one frequency sample.
  • Narrow-band interference does typically not, however, concentrate exactly point by point on one frequency, but may be slightly more widely spread. For instance, the bandwidth of an AM radio transmission including sidebands may be 9 kHz.
  • the interference may be spread in the area of several frequency samples. Because of this, it may be advantageous that the frequency range to be nulled comprises not only the frequency sample in which the interference shows, i.e. the cut level is exceeded, but also one or more neighbouring frequency samples depending on the sampling frequency being used. For instance, at a 16 MHz sampling frequency and with a 960,000-point sample string, the width of one frequency sample corresponds to approximately 16.7 Hz. If the width of the interference to be removed is 9 kHz, the number of frequency samples to be nulled is 540, i.e. 270 samples on both sides of the sample exceeding the cut level. The result is the corrected spectrum G(j ⁇ ) 1 .
  • Figure 6 shows an adjusted amplitude spectrum
  • the amplitude values exceeding the cut level 61 are also nulled 106 in the adjusted amplitude spectrum, in which case block 107 uses the corrected adjusted amplitude spectrum
  • the correction of the adjusted amplitude spectrum alters (makes smaller) its average and standard deviation, and the definition of the second cut level can be made more accurately in the second cutting. If more than two cuttings were used, the amplitude spectrum used in defining 105 and 107 the cut level would also be corrected correspondingly with each spectrum correction 106.
  • Figure 9 shows the corrected signal g(t) 2 in a time domain, obtained by transforming 109 the twice corrected 106 and 108 spectrum, i.e. signal, G(j ⁇ ) 2 from the frequency domain back to the time domain.
  • the narrow-band interference remaining in the signal are considerably lower than the background noise, so distinguishing interference in the time domain ( Figure 9) is almost impossible.
  • a method of eliminating asynchronous impulse interference is based on the fact that the typical properties of a partial discharge pulse and asynchronous interfering pulses differ from each other enough to make distinguishing them possible. Partial discharge pulses occur in pulse groups in areas of cycle depending on the discharge type, and there is deviation in the location and amplitude of a single pulse, whereas asynchronous impulse interference occurs at almost equal intervals, at nearly a constant amplitude and during the entire cycle.
  • the method uses pulse parameters obtained by finding the pulses occurring in a partial discharge measuring signal and calculating and storing the following information on them: amplitude, starting phase angle, cycle number, rise time, fall time, width and area of the pulse.
  • amplitude was scaled to between 0 and 128. No limits are set to the number of pulses.
  • the flow chart of the method is shown in Figure 11.
  • the pulses occurring in the measuring signal are found and pulse parameters (e.g. amplitude, phase angle and cycle number) are defined for the pulses.
  • Figure 14 shows a partial discharge measuring signal. The peaks of the pulses found in the signal have been circled. The pulse parameters of the found pulses are used as the starting data.
  • Figure 15 shows the peaks of the found pulses in time order. In Figures 14 and 15, the X axis represents the samples, and thus in the case described herein, one sample corresponds to 62.5 ns.
  • Figure 15 also shows a 50 Hz sine wave to illustrate the location of the pulses in the phase voltage more clearly.
  • the string of asynchronous interfering pulses is shown in Figure 15 as a string in the amplitude area 20 to 25.
  • the amplitude area (0 to 128) is preferably divided (step 112) into 41 examination periods, as shown in Figure 12.
  • the amplitude deviation i.e. length of the period, is preferably ⁇ 3.
  • the periods are examined 113 one by one and only the impulses, whose amplitude is within the period, are examined at each time.
  • the amplitude area is divided into smaller examination periods so as to be able to examine the time difference between pulses of certain amplitudes.
  • the time difference matrix contains the time difference (i.e. distance) of each pulse to all pulses preceding it within in the same examination period.
  • a time difference distribution is created 114.
  • the distribution is formed in the range of 0.04 to 4 ms, for instance.
  • the distribution is preferably divided into 91 logarithmic blocks so that the step from one block to another is 5.2%.
  • the value time difference/20 is added to the distribution and, on both sides of the time difference, the value (time difference/20)2.
  • Figure 16 shows the division of the blocks of the time difference distribution. The size of the blocks increases from left to right.
  • the distribution is formed of the time difference between consecutive pulses in the same examination period (i.e. the time difference of the pulse to the pulse preceding it). In practice, the values from which the distribution is formed are immediately above the diagonal axis of the time difference matrix.
  • Figure 17 shows a time difference distribution formed for the exemplary partial discharge signal mentioned earlier in an examination period of 21 ⁇ 3.
  • the distribution shows a strong peak at approximately 1 ms and the peak is higher than the value 1 set as the threshold, and consequently on the basis of the distribution, it can be noted that the signal has asynchronous impulse interference at 1 ms (16,000 samples) intervals.
  • the value time difference/20 is a density index.
  • the height of the peak in the distribution shown in Figure 17 is approximately 1.8.
  • the peak value for three cycles is 3, so the peak shows that over a half of the pulse pairs occurring at 1 ms intervals have been detected.
  • the time difference of consecutive pulses only is stored in the distribution.
  • the value time difference/20 means that the value 1 will be obtained for the distribution, if the cycle has 20 pulses at 1 ms intervals or 100 pulses at 0.2 ms intervals.
  • the distribution thus shows the time difference at which interval equidistant pulses occur proportionally the most.
  • the maximum of the time difference distribution is found 115 and the time difference with which it is realised. If the peak value of the time difference distribution is higher than a pre-set threshold value, pulses repeating at equal intervals in the time difference matrix are searched for. A three-cycle long measurement uses the value 1 as the threshold value. This requires that when forming the time difference distribution, at least every third of the repeating pulse pairs have been detected. The threshold value should not be set too high, because it may happen that there are both partial discharge pulses and interfering pulses in the same examination period.
  • time difference is, however, so high that partial discharges do not exceed it. Not even the pulse strings of a corona discharge are "long" enough to be interpreted as interference.
  • the value time difference/20 scales the distribution so that the highest possible value of the distribution corresponds to the number of cycles. If there were samples from one network cycle only, the peak value would be 1. In the example, three network cycle long packets are examined, and the peak value of the distribution is 3.
  • Finding 116 pulses that repeat at equal intervals in the time difference matrix is done as follows.
  • the search is started using pulse 1 as the first pulse. From the first horizontal line a value is searched which is within the range: k ⁇ AE - 0.135 ⁇ AE ⁇ Value ⁇ k ⁇ AE + 0.135 ⁇ AE, wherein AE is the time difference between repeating pulses, obtained from the time difference distribution and k is 1, 2, 3...
  • Figure 20 shows the values of a time difference matrix for the first 31 pulses.
  • the time difference between repeating pulses is 1 ms in the example. In this case, the value meeting the conditions is found on line 1, location 22 of the matrix.
  • the pulse index is stored so that it can be marked as an interfering pulse, if the found pulse string meets the conditions set for it.
  • the pulse search is now continued from line 22. A value meeting the conditions cannot be found in this table and it can be noted later that several pulses are missing from the found pulse string, i.e. it is not uniform enough to be an interfering pulse string.
  • the search is then started from the beginning using pulse 2 as the first pulse. A value meeting the conditions is searched for on the 2 nd vertical line. It is found at location 3.
  • the pulse index is stored and the search is continued on line 3.
  • a pulse string is searched for from the time difference matrix, whose time difference between pulses is k x 1 ms ⁇ 0.135 ms.
  • the interfering pulses found in the exemplary signal are circled in Figure 18.
  • the pulse string is sufficiently uniform.
  • the condition is preferably that at least half of the pulses have been found. If the pulse string is sufficiently uniform, the pulses in it are marked 117 as interfering pulses.
  • Figure 19 shows an exemplary signal from which all detected interfering pulses have been removed.
  • the following parameters can be changed in the method: amplitude deviation, number of examination periods, time difference deviation and number of searched asynchronous interfering pulses.
  • the time difference matrix can be replaced by calculating only the values immediately above its diagonal, i.e. by defining the time difference of consecutive pulses. The time difference between two pulses can be calculated using these values, but the calculation becomes more complicated.
  • a method of eliminating synchronous impulse interference is based on the fact that the typical properties of a partial discharge pulse and synchronous interfering pulses differ from each other enough to make distinguishing them possible. Partial discharge pulses occur in pulse groups in areas of cycle depending on the discharge type, but there is deviation in the location and amplitude of a single pulse, whereas synchronous impulse interference occurs at nearly the same phase angle and at nearly constant amplitude.
  • the method preferably uses pulse parameters obtained by finding the pulses occurring in a partial discharge measuring signal and calculating and storing 120 the amplitude, starting phase angle, cycle number, rise time, fall time, width and area of the pulse.
  • the amplitude of the pulses was scaled to between 0 and 128.
  • the number of synchronous interfering pulses during one cycle is assumed to vary case by case.
  • the flow chart of the method is shown in Figure 21.
  • the amplitude area (0 to 128) is divided (step 201) into smaller examination periods like in connection with asynchronous impulse interference elimination.
  • the amplitude area is, for instance, divided into 20 partly overlapping examination periods and the width of a period is ⁇ 0.055 x the amplitude of the largest pulse in the pulse series.
  • the examination periods are examined 202 one at a time and only the pulses are examined, whose amplitude is within the period.
  • a phase angle distribution of the pulses is formed 203 by checking all pulses in the same examination period and forming the distribution from their starting phase angles.
  • the distribution is divided into 180 blocks, for instance, in relation to the phase, i.e. the width of one block corresponds to 2°.
  • Forming the distribution is started from the first measured cycle and the first block of the distribution, i.e. phase angles 0 to 2°.
  • the value 1 is added to the first block of the distribution, if in the first cycle, only one pulse occurs in said block and no pulses occur within two blocks of the examined block, i.e. phase angles 2 to 6°. If the fifth block, for instance, were examined, i.e.
  • phase angles 8 to 10° there should be no pulses at the phase angles 4 to 8° and 10 to 14°. If there are more than one pulse in the block or if there are pulses in the neighbouring blocks, the value of the distribution is not changed. This action endeavours to prevent the elimination of partial discharge pulses. Even though partial discharge pulses are quite irregular, it is possible that, in the middle of pulse groups, pulses occur so densely that some of them may be identified as synchronous impulse interference. The drawback in this is that synchronous interfering pulses occurring at the same phase angles as partial discharges cannot necessarily be identified. The obtained distribution is rounded by summing the values of the neighbouring blocks to each block.
  • the phase angle distribution is normed by dividing its values by the number of cycles. Any values higher than 0.4 are interpreted 204 as peaks of the phase angle distribution. If peaks are detected, the pulses are re-examined and the pulses occurring at phase angles corresponding to the peaks are marked 205 as interfering pulses. The marking as interfering pulses is, however, done so that only one pulse per cycle is marked as an interfering pulse in one phase window. Thus, if one phase window has both a partial discharge pulse and an interfering pulse, at least one of them will be analysed.
  • the pulse parameters of the pulses that are marked as interfering pulses are removed 207 from the database.
  • the following parameters can be changed: amplitude deviation, number of examination periods and threshold value of peak search.
  • the analysis limit is a thousand pulses, for instance.
  • a collected group of for instance a thousand pulses is in the following called a sample.
  • One or more characteristic parameters are defined 301 of the collected pulses.
  • the characteristic parameters may correspond to the pulse parameters defined 400 earlier from the pulses, in which case the values of the corresponding pulse parameters are defined as the values of the characteristic parameters.
  • the following three characteristic parameters are defined:
  • a rise time histogram is first formed.
  • Figure 24 shows an example of a rise time histogram.
  • the rise time is shown on the horizontal axis of the rise time histogram divided into certain intervals and the number of pulses in each interval is shown on the vertical axis.
  • the rise time histogram is preferably filtered with a FIR (finite impulse response) filter of the eighth order, whose barrier frequency is suitably selected.
  • the barrier frequency can be 0.1 x f N , for instance.
  • the frequency f N is a Nyquist frequency, i.e. half of the sampling frequency.
  • the histogram is processed as a sample string whose sampling frequency is 2 Hz, for instance, i.e. f N becomes 1 Hz and the barrier frequency of the filter 0.1 Hz.
  • Figure 25 shows an example of a filtered rise time histogram.
  • the histogram is divided into classes on the basis of minimum local values between peaks in such a manner, for instance, that first the highest value of the filtered histogram is searched for, which in Figure 25 is at 50 ns. After this, peaks which are higher than 0.5 x the highest value, for instance, are searched for. In the case of Figure 25, a peak meeting the condition is at approximately 150 ns.
  • the smallest value between peaks is defined as the minimum value between peaks, this being at 100 ns in Figure 25.
  • the highest value of the histogram is thus set as the first peak and, as the next peaks, the ones whose height is at least the set limit value (e.g. 0.2 to 0.5 x the highest peak).
  • a peak refers to a value (peak) which is higher than the values surrounding it.
  • a peak can also extend over more than one interval.
  • Pulse width histograms are then formed of each pulse group formed on the basis of the rise and fall time histograms.
  • the histograms are filtered with the above-mentioned FIR filter and divided into groups on the basis of the minimum values between peaks, as described above.
  • Separate pulse groups are formed of pulses belonging to different groups.
  • the pulse groups formed on the basis of the rise and fall time histograms are thus divided into subgroups, if the groups have pulses with different pulse widths.
  • the size of the interval in the histograms is bound to the sampling interval t s (e.g. 62.5 ns) of the measurement system so that the interval of the rise time histogram is preferably 0.4 x t s (e.g.
  • each pulse group only contains the pulses of one or more discharge sources located at a certain distance from the measuring point.
  • the number of pulse groups is obtained 305 as the number of discharge sources at different distances from the measuring point.
  • partial discharge sources of different type located at the same distance are distinguished, if their pulse forms differ substantially from each other. Otherwise, partial discharge sources located at the same distance can be identified and distinguished 305 either manually (by examining the discharge pulse groups visually) or automatically by using an identification method of partial discharge types.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Testing Relating To Insulation (AREA)
EP00660185A 1999-10-19 2000-10-11 Procédé et dispositif de détermination du nombre de sources de décharges partielles Expired - Lifetime EP1094325B1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
FI992260A FI107081B (fi) 1999-10-19 1999-10-19 Menetelmä ja järjestely osittaispurkauslähteiden lukumäärän selvittämiseksi
FI992260 1999-10-19

Publications (3)

Publication Number Publication Date
EP1094325A2 true EP1094325A2 (fr) 2001-04-25
EP1094325A3 EP1094325A3 (fr) 2004-08-11
EP1094325B1 EP1094325B1 (fr) 2006-06-14

Family

ID=8555472

Family Applications (1)

Application Number Title Priority Date Filing Date
EP00660185A Expired - Lifetime EP1094325B1 (fr) 1999-10-19 2000-10-11 Procédé et dispositif de détermination du nombre de sources de décharges partielles

Country Status (4)

Country Link
US (1) US6507181B1 (fr)
EP (1) EP1094325B1 (fr)
DE (1) DE60028684T2 (fr)
FI (1) FI107081B (fr)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1439396A2 (fr) * 2003-01-15 2004-07-21 Alstom Technology Ltd Procédé d'analyse et de surveillance du comportement de la décharge partielle d'une installation électrique
CN104614639A (zh) * 2010-03-05 2015-05-13 环境股份(有限)公司 评估电力线上的噪声和过流
CN106569095A (zh) * 2016-11-09 2017-04-19 于祥茹 一种基于加权平均依赖分类器的电网故障诊断***

Families Citing this family (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CA2508428A1 (fr) * 2005-05-20 2006-11-20 Hydro-Quebec Detection, localisation et interpretation de decharge partielle
US7532012B2 (en) * 2006-07-07 2009-05-12 Ambient Corporation Detection and monitoring of partial discharge of a power line
CA2573711A1 (fr) 2007-01-10 2008-07-10 Hydro Quebec Controle gamma de l'equivalent en eau de la neige et de l'humidite du sol
CN102517280A (zh) * 2007-01-31 2012-06-27 菲尼克斯股份有限公司 用于提高表达的细菌前导序列
US7676333B2 (en) * 2007-11-06 2010-03-09 General Electric Company Method and apparatus for analyzing partial discharges in electrical devices
US8143899B2 (en) * 2008-04-01 2012-03-27 General Electric Company Method and apparatus for detecting partial discharges in electrical systems
ES2545098T3 (es) * 2010-07-26 2015-09-08 Prysmian S.P.A. Aparato y procedimiento de monitorización de un sistema de transmisión de energía eléctrica mediante análisis de descargas parciales
DE102011082866A1 (de) * 2011-09-16 2013-03-21 Olaf Rudolph Verfahren zum Erkennen einer oder meherer gleichzeitig auftretender Teilenladungs-Quellen
US9557365B2 (en) 2013-03-14 2017-01-31 Hubbell Incorporated Apparatuses, systems and methods for detecting corona
CN105122072B (zh) * 2013-04-22 2018-06-22 三菱电机株式会社 电力设备的部分放电检测方法以及部分放电检测装置
CN103630584B (zh) * 2013-10-16 2015-07-15 浙江工商大学 一种脉冲弛豫信号时域拐点提取方法
US20150142344A1 (en) * 2013-10-18 2015-05-21 Utilx Corporation Method and apparatus for measuring partial discharge charge value in frequency domain
EP2947668B1 (fr) * 2014-05-21 2023-08-02 Siemens Energy Global GmbH & Co. KG Circuit pour une installation de contrôle haute tension
KR101787901B1 (ko) * 2016-06-14 2017-11-15 엘에스산전 주식회사 전력설비 진단장치
CA3007729A1 (fr) 2017-06-12 2018-12-12 Vibrosystm Inc. Methode de surveillance de decharges partielles dans une machine electrique a haute tension, et cable de connexion associe
RU2724455C1 (ru) * 2017-06-14 2020-06-23 Мицубиси Электрик Корпорейшн Оборудование диагностики ухудшения характеристик вследствие старения и способ диагностики ухудшения характеристик вследствие старения
JP6730354B2 (ja) * 2018-03-23 2020-07-29 ファナック株式会社 入力ミス検出装置
JP7362205B2 (ja) * 2019-08-26 2023-10-17 日東工業株式会社 放電検出装置
CN111413596A (zh) * 2020-03-06 2020-07-14 南京南瑞继保工程技术有限公司 电缆局部放电定位方法及装置、电子设备及定位***

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4710705A (en) * 1984-09-28 1987-12-01 Fuji Electric Company, Ltd. Noise removal circuit for use in a partial discharge measuring device of a high voltage apparatus
US5475312A (en) * 1994-06-07 1995-12-12 Iris Power Engineering Inc. Method and device for distinguishing between partial discharge and electrical noise
WO1998047009A2 (fr) * 1997-04-11 1998-10-22 General Electric Company Methodes statistiques d'analyse de configuration de mesure de decharges partielles dans une isolation haute tension

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2997571B2 (ja) * 1991-07-10 2000-01-11 三菱電機株式会社 監視診断装置
JP3187642B2 (ja) * 1994-02-25 2001-07-11 関西電力株式会社 電気機器の異常検出方法及び回転電機の異常検出装置

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4710705A (en) * 1984-09-28 1987-12-01 Fuji Electric Company, Ltd. Noise removal circuit for use in a partial discharge measuring device of a high voltage apparatus
US5475312A (en) * 1994-06-07 1995-12-12 Iris Power Engineering Inc. Method and device for distinguishing between partial discharge and electrical noise
WO1998047009A2 (fr) * 1997-04-11 1998-10-22 General Electric Company Methodes statistiques d'analyse de configuration de mesure de decharges partielles dans une isolation haute tension

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
BOZZO R ET AL: "BP networks for partial discharge analysis and dielectric classification" NEURAL NETWORKS,1997., INTERNATIONAL CONFERENCE ON HOUSTON, TX, USA 9-12 JUNE 1997, NEW YORK, NY, USA,IEEE, US, 9 June 1997 (1997-06-09), pages 37-42, XP010238788 ISBN: 0-7803-4122-8 *
PAKONEN P ET AL: "The identification of discharge type by radio interference measurements" NORDIC INSULATION SYMPOSIUM NORD-IS 92, VASTERAS, SWEDEN, 15-17 JUNE 1992, pages 7.8/1-10, XP008022480 1992, Vasteras, Sweden, Nygrens Ljuakopla AB, Sweden *

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1439396A2 (fr) * 2003-01-15 2004-07-21 Alstom Technology Ltd Procédé d'analyse et de surveillance du comportement de la décharge partielle d'une installation électrique
EP1439396A3 (fr) * 2003-01-15 2005-01-19 Alstom Technology Ltd Procédé d'analyse et de surveillance du comportement de la décharge partielle d'une installation électrique
US7076404B2 (en) 2003-01-15 2006-07-11 Alstom Technology Ltd. Method and apparatus for the analysis and monitoring of the partial discharge behavior of an electrical operating device
CH696646A5 (de) * 2003-01-15 2007-08-31 Alstom Technology Ltd Verfahren zur Analyse und/oder Überwachung des Teilentladungsverhaltens eines elektrischen Betriebsmittels.
CN104614639A (zh) * 2010-03-05 2015-05-13 环境股份(有限)公司 评估电力线上的噪声和过流
CN104614639B (zh) * 2010-03-05 2018-02-06 瑞典爱立信有限公司 评估电力线上的噪声和过流
CN106569095A (zh) * 2016-11-09 2017-04-19 于祥茹 一种基于加权平均依赖分类器的电网故障诊断***
CN106569095B (zh) * 2016-11-09 2019-07-26 于祥茹 一种基于加权平均依赖分类器的电网故障诊断***

Also Published As

Publication number Publication date
EP1094325B1 (fr) 2006-06-14
DE60028684T2 (de) 2006-10-12
US6507181B1 (en) 2003-01-14
EP1094325A3 (fr) 2004-08-11
FI107081B (fi) 2001-05-31
DE60028684D1 (de) 2006-07-27

Similar Documents

Publication Publication Date Title
US6507181B1 (en) Arrangement and method for finding out the number of sources of partial discharges
EP1094323B1 (fr) Procédé et appareil pour identifier la cause d'une décharge partielle
EP1094324B1 (fr) Procédé et appareil pour localiser la source de décharges partielles
Zhang et al. A novel wavelet transform technique for on-line partial discharge measurements. 1. WT de-noising algorithm
RU2532142C2 (ru) Способ и система мониторинга частичных разрядов
CN104459779B (zh) 一种异常地震道的自动识别的方法和装置
US9989581B2 (en) Method and device for locating partial discharges in electric cables
US8447540B2 (en) Method for detecting, identifying and locating partial discharges occurring in a discharge site along an electric apparatus
KR20090045154A (ko) 전력선의 부분 방전의 검출 및 감시
CN107167716A (zh) 一种局部放电缺陷类型识别方法及装置
KR102393662B1 (ko) 전력 기기 부분방전 진단 시스템 및 방법 및 상기 방법을 실행시키기 위한 컴퓨터 판독 가능한 프로그램을 기록한 기록 매체
CN105091732A (zh) 一种变压器绕组变形的检测方法及***
JP7157261B2 (ja) 部分放電監視装置及び部分放電監視方法
JP5539762B2 (ja) 避雷装置の故障判定方法
CN104865508A (zh) 基于数据分组量化的局部放电识别方法
CN117434396A (zh) 一种变压器套管末屏在线监测***及其方法
JP6215992B2 (ja) 電力ケーブル絶縁劣化位置推定方法及び推定システム
US20130307524A1 (en) Inferring the periodicity of discrete signals
CN116593842A (zh) 一种宽频带、带滤波功能的电缆局部放电检测***及方法
US6853933B2 (en) Method of identifying spectral impulses for Rj Dj separation
CN114924328B (zh) 一种带垂直磁场参考道的城市人工源电磁勘探方法及***
CN111886510B (zh) 量化包括高斯和有界分量的随机定时抖动
CN107728207B (zh) 地震勘探单炮记录声波识别方法和单炮质量检验方法
CN116106368A (zh) 一种血细胞电压信号的处理方法和***
WO2022152339A1 (fr) Procédé d'extraction de signal et dispositif pour sa mise en œuvre

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

AK Designated contracting states

Kind code of ref document: A2

Designated state(s): AT BE CH CY DE DK ES FI FR GB GR IE IT LI LU MC NL PT SE

AX Request for extension of the european patent

Free format text: AL;LT;LV;MK;RO;SI

PUAL Search report despatched

Free format text: ORIGINAL CODE: 0009013

AK Designated contracting states

Kind code of ref document: A3

Designated state(s): AT BE CH CY DE DK ES FI FR GB GR IE IT LI LU MC NL PT SE

AX Request for extension of the european patent

Extension state: AL LT LV MK RO SI

17P Request for examination filed

Effective date: 20041230

RAP1 Party data changed (applicant data changed or rights of an application transferred)

Owner name: ABB OY

AKX Designation fees paid

Designated state(s): DE FR GB

17Q First examination report despatched

Effective date: 20050407

GRAP Despatch of communication of intention to grant a patent

Free format text: ORIGINAL CODE: EPIDOSNIGR1

GRAS Grant fee paid

Free format text: ORIGINAL CODE: EPIDOSNIGR3

RIN1 Information on inventor provided before grant (corrected)

Inventor name: PAKONEN, PERTTI

Inventor name: LATVA-PUKKILA, VESA

Inventor name: BJOERKVIST, MATS

RAP1 Party data changed (applicant data changed or rights of an application transferred)

Owner name: ABB OY

RIN1 Information on inventor provided before grant (corrected)

Inventor name: BJOERKQVIST, MATS

Inventor name: PAKONEN, PERTTI

Inventor name: LATVA-PUKKILA, VESA

GRAA (expected) grant

Free format text: ORIGINAL CODE: 0009210

AK Designated contracting states

Kind code of ref document: B1

Designated state(s): DE FR GB

REG Reference to a national code

Ref country code: GB

Ref legal event code: FG4D

REF Corresponds to:

Ref document number: 60028684

Country of ref document: DE

Date of ref document: 20060727

Kind code of ref document: P

ET Fr: translation filed
PLBE No opposition filed within time limit

Free format text: ORIGINAL CODE: 0009261

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: NO OPPOSITION FILED WITHIN TIME LIMIT

26N No opposition filed

Effective date: 20070315

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: FR

Payment date: 20131022

Year of fee payment: 14

Ref country code: GB

Payment date: 20131021

Year of fee payment: 14

Ref country code: DE

Payment date: 20131021

Year of fee payment: 14

REG Reference to a national code

Ref country code: DE

Ref legal event code: R119

Ref document number: 60028684

Country of ref document: DE

GBPC Gb: european patent ceased through non-payment of renewal fee

Effective date: 20141011

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: GB

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20141011

Ref country code: DE

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20150501

REG Reference to a national code

Ref country code: FR

Ref legal event code: ST

Effective date: 20150630

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: FR

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20141031